Temporarily disable workflow testing while investigating actions test failure

[fynyky]

No description provided.

[claude]

Claude encountered an error after 1m 50s —— View job

---

Claude Code Review

• Read PR diff and changed files
• Analyze workflow changes
• Provide review feedback

---

Summary

This PR comments out the test workflow (.github/workflows/test.yml) entirely and removes the npm test step from the publish workflow (.github/workflows/publish.yml) to unblock things while the Actions test failure is investigated. Mechanically, it does what it says — but as a "temporary" change it introduces some risk and could be done more cleanly.

Note: This PR is already merged, so this is review feedback for the historical record and for the follow-up that re-enables tests.

---

Concerns

1. npm publish now runs without any test or lint gate — .github/workflows/publish.yml:21-25

Commenting out - run: npm test in the publish workflow means anything pushed to release will get shipped to npm without:

• the test suite passing, and
• linting (standard), which is invoked via the pretest script (package.json:14) — that gate is now silently disabled too.

For a public package (@fynyky/elemental) this is the most consequential change here. Even while the test workflow is broken, you probably want some signal before publishing. Options:

• Run lint directly in the publish job (- run: npm run lint) — fast and doesn't depend on Playwright.
• Keep npm test and add continue-on-error: true so failures are visible but non-blocking.
• Hold off on merges to release until tests are re-enabled (procedural rather than CI).

Fix this →

2. "Temporary" with no tracking mechanism

The PR title and commit message both say "temporarily" / "while investigating," but there's nothing in the diff that will remind anyone to re-enable this:

• No TODO: / FIXME: comment with a link to a tracking issue.
• No linked issue in the PR description (the body is empty).
• No date or condition for re-enablement.

A # TODO(#NN): re-enable after fixing <root cause> comment at the top of each disabled block would make this visible to anyone who later opens the file. Consider also opening a tracking issue and linking the failing CI run (https://github.com/fynyky/elemental/actions/...) so the investigation context isn't lost.

3. Whole-file comment-out is messier than alternatives — .github/workflows/test.yml

Commenting out every line of a workflow file leaves a "ghost" file in the repo that's easy to overlook. Cleaner alternatives that signal intent better:

• Disable the job with if: false — keeps the file valid YAML, makes the intent explicit, and re-enabling is a one-line change:

  jobs:
    test:
      if: false  # TODO(#NN): re-enable once <root cause> is fixed
      runs-on: ubuntu-latest
      ...

• Restrict triggers to workflow_dispatch — lets you still run the workflow manually while you investigate, which is useful for iterating on the fix.
• Delete the file — git history preserves it; restoring is git revert. Cleaner than 21 lines of commented YAML.

Of these, if: false (option 1) is the lowest-friction and is what I'd recommend for a self-described "temporary" disable.

4. Pre-existing minor issues that this PR could have cleaned up

• .github/workflows/publish.yml has no trailing newline (the diff shows \ No newline at end of file). Worth fixing while you're in there.
• branches: master in both files is the string form; the list form branches: [master] is more conventional and avoids ambiguity with branches-ignore semantics. Pre-existing, not introduced here.

---

What looks fine

• The mechanical change is consistent across both files — both the standalone test.yml and the test step inside publish.yml are disabled together, which is the right scope. Disabling one without the other would have left a confusing half-state.
• No application code is touched — the blast radius is contained to CI configuration.
• Single, focused commit with a clear message.

---

Recommended follow-up

1. Open an issue tracking the actual test failure (with a link to the failing run) so the investigation has a home.
2. Replace the commented-out blocks with if: false + a TODO(#issue) comment so the disable is visible and trivially reversible.
3. Add a npm run lint step to the publish workflow so npm releases keep some gate.
4. Once the underlying Playwright/test failure is fixed, the revert is one PR.
   | Branch: claude/pr-72-20260529-1350